Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
In a recent interview with China’s state-run People’s Daily, Huawei’s founder Ren Zhengfei downplayed his company’s technological capabilities amidst trade talks between Beijing and Washington. Ren stated that Huawei’s Ascend chip, a competitor to Nvidia’s products in China, still lags behind the US by one generation. He also mentioned that the US has exaggerated Huawei’s capabilities and that the company is not as strong as portrayed.
These comments come as Nvidia CEO Jensen Huang has expressed concerns about Huawei’s progress with its artificial intelligence chips, highlighting that restrictions on Nvidia selling to China have created a formidable competitor in the AI technology space.
The US and China have engaged in new trade talks that include discussions on export controls on key technologies. The recent restrictions imposed by Beijing on critical rare earths and minerals used in manufacturing have brought the issue of export controls to the forefront of these discussions.
Despite the ban on Nvidia shipments to China, Huawei has seen increased demand for its Ascend chips as Chinese tech giants look to adopt the company’s technology. However, most Chinese AI companies, like DeepSeek, still rely on Nvidia chips for training large language models, while using domestic alternatives for less complex tasks.
Analysts and researchers have raised concerns about technical glitches in using Huawei’s chips for training language models. Ren hinted at advancements in resolving these issues, mentioning that Huawei could compensate for performance issues through cluster computing.
Huawei’s CloudMatrix 384 AI server, which connects 384 AI processors using optical technology, is a key component of the company’s strategy to compete with Nvidia. Several Huawei customers are currently testing the server, working with engineers to address issues related to heat generation and weight.
Creating an ecosystem of developers to build AI models remains a challenge for Huawei, with Nvidia’s software platform Cuda being a preferred choice for developers. Ren mentioned that Huawei is investing significantly in research and development, with a focus on basic research to make breakthrough discoveries.
Ren highlighted China’s advantages in building technological capabilities, citing the country’s advanced power generation and telecom infrastructure as key factors for AI development.
Overall, Huawei continues to make strides in the AI technology space, with ongoing efforts to compete with established players like Nvidia. The company’s investments in research and development signal a commitment to innovation and advancing technological capabilities.
